ChanServ changed the topic of #asahi-gpu to: Asahi Linux: porting Linux to Apple Silicon macs | GPU / 3D graphics stack black-box RE and development (NO binary reversing) | Keep things on topic | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Logs: https://alx.sh/l/asahi-gpu
<alyssa> lina: "ail: Do not unconditionally page-align sizes" is breaking t-rex for me
<alyssa> unclear why
<alyssa> if I functionally revert it, works again
<alyssa> I am unsure where you got the logic from
<alyssa> if I align for uncompressed but not for twiddled, it seems happy
<alyssa> although my unti test failing no
<alyssa> ok did the obvious fix and it's happy
<alyssa> Worth noting that t-rex uses (IIRC) ETC2 or ASTC cube maps
<alyssa> which are uncompressed twiddled and the alignment matters
<alyssa> splat for you
<alyssa> it should not be possible to provoke an rtkit crash from userspace so this makes me profoundly uncomfortable
<alyssa> also just hit the oom killer yikes
<alyssa> leaking staging resources, just fixed that locally
* alyssa fixes stuff locally
<Tramtrist> the only way to fix things really
<alyssa> Tramtrist: i've fixed plenty of stuff remotely :-p
<alyssa> cube maps + compression is botched because we can't do linear staging 2D Array resources
<alyssa> i've just disabled non-2D compression for now because that's nontrivial to workaround (though certainly doable)
<alyssa> similarly, compression to non-renderable formats doesn't work
<alyssa> in principle all formats are renderable if we want them to be, but..
<alyssa> it's worth seeing if Apple uses this compression for funny stuff that shouldn't be renderable :p
le0n has joined #asahi-gpu
ahubcap has joined #asahi-gpu
<alyssa> ok, bunch of fixes and we're back to no gles2 regressions
<alyssa> gles3 is still going but seems about as broken as usual
millenialhacker has joined #asahi-gpu
le0n_ has quit [Ping timeout: 480 seconds]
ahubcap_ has joined #asahi-gpu
millenialhacker has quit [Ping timeout: 480 seconds]
ahubcap has quit [Ping timeout: 480 seconds]
ahubcap_ has quit [Quit: Leaving...]
millenialhacker has joined #asahi-gpu
millenialhacker has quit [Ping timeout: 480 seconds]
millenialhacker has joined #asahi-gpu
millenialhacker has quit [Ping timeout: 480 seconds]
<alyssa> Pushed a new version to https://gitlab.freedesktop.org/asahi/mesa
<alyssa> the latter should help performance noticeably, and it's pretty well tested now
<alyssa> I tested my current set of { GNOME, neverball, supertuxkart, quake, xonotic }
<alyssa> but maybe other people will find other issues
<alyssa> jannau: yuka: You may wish to upgrade ^^
<alyssa> re performance, I have a LOT more patches where that came from but letting them trickle out to give time to test everything
<alyssa> [My unstable Linux integration branch has asahi/main + NIR UBO/VBO lowering + jmp_exec_none + framebuffer compression + DISCARD_WHOLE_RESOURCE]
<alyssa> (Seeing as 2 of those were written today and 1 of those I found big issues with as recently as this afternoon, they'll be a bit longer. But they're coming)
CatDentures has quit [Quit: Konversation terminated!]
trsohmers has quit [Ping timeout: 480 seconds]
trsohmers has joined #asahi-gpu
SSJ_GZ has joined #asahi-gpu
millenialhacker has joined #asahi-gpu
millenialhacker has quit [Ping timeout: 480 seconds]
millenialhacker has joined #asahi-gpu
<jannau> alyssa: works and is faster, thanks
millenialhacker has quit [Ping timeout: 480 seconds]
zimsneexh has joined #asahi-gpu
zimsneexh has quit []
zimsneexh has joined #asahi-gpu
zimsneexh has quit [Quit: zimsneexh]
zimsneexh_ has joined #asahi-gpu
zimsneexh_ is now known as zimsneexh
millenialhacker has joined #asahi-gpu
capta1nt0ad has joined #asahi-gpu
millenialhacker has quit [Ping timeout: 480 seconds]
millenialhacker has joined #asahi-gpu
millenialhacker has quit [Ping timeout: 480 seconds]
zimsneexh has quit [Ping timeout: 480 seconds]
millenialhacker has joined #asahi-gpu
millenialhacker has quit [Ping timeout: 480 seconds]
millenialhacker has joined #asahi-gpu
millenialhacker has quit [Ping timeout: 480 seconds]
millenialhacker has joined #asahi-gpu
millenia_ has joined #asahi-gpu
millenialhacker has quit [Ping timeout: 480 seconds]
millenia_ has quit []
MajorBiscuit has joined #asahi-gpu
Major_Biscuit has joined #asahi-gpu
MajorBiscuit has quit [Ping timeout: 480 seconds]
a_plastic_bag has joined #asahi-gpu
a_plastic_bag has quit [Quit: WeeChat 3.7.1]
Major_Biscuit has quit [Ping timeout: 480 seconds]
<mps> hi. newbie question about building kernel asahi gpu with rust. build failed. is it ok to post paste link with build log here?
bisko has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Dcow> mps: pm'd you
capta1nt0ad has quit [Quit: Konversation terminated!]
capta1nt0ad has joined #asahi-gpu
cylm has joined #asahi-gpu
<mps> nice, Dcow helped me to solve build problem. thank you Dcow
zimsneexh has joined #asahi-gpu
<Dcow> np
<Dcow> I love procrastinate from $work sometimes
zimsneexh has quit [Quit: zimsneexh]
<mps> now another question, does xorg works with gpu or wayland is only option
<Dcow> wayland is preffered
<_jannau_> but there could be an incompatibility with xorg/modeset
<mps> I've got this fatal error when tried to start X '(EE) Cannot run in framebuffer mode. Please specify busIDs for all framebuffer devices'
Major_Biscuit has joined #asahi-gpu
chadmed_ has joined #asahi-gpu
<chadmed_> mps: wayland and xwayland both work on my end, but same as you no "real" x11
<mps> thanks all for info
qdot has quit [Ping timeout: 480 seconds]
bisko has joined #asahi-gpu
<alyssa> jannau: +1
<alyssa> i saw real x11 working once and have since forgotten how i did it :p
bisko has quit [Ping timeout: 480 seconds]
qdot has joined #asahi-gpu
Jondus has joined #asahi-gpu
<Jondus> Hello :)
<Jondus> I got Asahi Mesa fully built. So does that mean it setups Wayland to use Mesa or do I need to do that manually?
<Jondus> I ask because my Wayland app is still not GPU accelerated. .. Maybe I need to reboot though lul
<_jannau_> how did you build it? if you want every application using it you have install it to replace the system mesa libraries
<Jondus> $ mkdir build $ cd build $ meson .. $ sudo ninja install
<Jondus> That should replace system mesa right?
alyssa has quit [Quit: leaving]
<_jannau_> not sure what mesa's default install path is
<Jondus> Ah.. Anyway you know of to force it to system path then?
<Dcow> Jondus: you need to enable gallium asahi
<Dcow> when build mesa
<Jondus> Dammit I knew something was off
<Dcow> + I assume you kernel already with DRM_ASAHI, right?
<Jondus> I'm using bleeding-edge
<Dcow> it's not enabled there
<Jondus> ugh. Can you tell me how to enable DRM_ASAHI then please
<Dcow> you need to prepare kernel yourself
<Jondus> Ah so I need to compile the kernel
<Jondus> got it
<Jondus> Dcow: Mind if I pop you a dm?
<Dcow> I have a meeting in few minutes, so probably won't answer
<Jondus> My apologies no problem
<Dcow> But general advice is read this chat logs and if you still have questions about building kernel process just wait until it reach -edge kernel =)
<Dcow> But if you have some specific questions pm is welcome, I just answer later
<Jondus> Pretty sure I just need to run make LLVM=1 DRM_ASAHI=1?
<Dcow> no
<Dcow> read about configuring kernel's
<Jondus> yep def in the wrong repo then
<Jondus> Anyway yah, I will do some more research
<Jondus> Thanks for your time
<Jondus> I was in the right repo but wrong command. I've never actually had to build a kernel like this
<Jondus> ffs.. It was make config
<Jondus> make menuconfig*
<marcan> please take build tutorials to #asahi-offtopic or PMs. this channel is for GPU development, not installation instructions.
<marcan> if you've never built a kernel before you probably shouldn't be trying out the GPU driver at this point.
<Jondus> I have build a kernel just this particular one. I will move to off-topic for further questions.
<marcan> you also need to rebuild m1n1 and update your DTs, which could leave your system unbootable.
<Jondus> just not this particular one*
<marcan> unless you know what you are doing and how to recover from that, you probably shouldn't
<Dcow> that's a strong point^
chadmed_ has quit [Remote host closed the connection]
Major_Biscuit has quit [Ping timeout: 480 seconds]
bisko has joined #asahi-gpu
gaminglt has joined #asahi-gpu
gaminglt has quit []
milan has joined #asahi-gpu
milan has quit []
Znullptr has joined #asahi-gpu
zalyx has quit [Quit: later alligator]
zalyx has joined #asahi-gpu
Cyrinux has quit [Remote host closed the connection]
Jondus has quit [Quit: Page closed]
Dcow has quit [Remote host closed the connection]
Dcow has joined #asahi-gpu
Dcow has quit []
Cyrinux has joined #asahi-gpu
Cyrinux has quit []
Cyrinux has joined #asahi-gpu
Major_Biscuit has joined #asahi-gpu
kenzie7 has quit []
kenzie7 has joined #asahi-gpu
kenzie7 has quit []
kenzie7 has joined #asahi-gpu
Cyrinux has quit []
Cyrinux has joined #asahi-gpu
bisko has quit [Read error: Connection reset by peer]
bisko has joined #asahi-gpu
Dcow has joined #asahi-gpu
erik3424 has joined #asahi-gpu
Major_Biscuit has quit [Ping timeout: 480 seconds]
<Dcow> alyssa: xonotic crash with asahi:main https://paste.gnome.org/QLosMwOdc
<Dcow> ../src/gallium/drivers/asahi/agx_state.c:1355
alyssa has joined #asahi-gpu
<alyssa> dcow: Odd... is that reproducible?
<Dcow> yes
<Dcow> set ultra settings, play ~1 minute, usually earlier (second map, idk if it matter)
<alyssa> oh I think I see the problem
<Dcow> xonotic-sdl
Cyrinux has quit [Remote host closed the connection]
<alyssa> try replace PIPE_MAX_SAMPLERS a few lines above with ctx->stage[shader].sampler_count
<alyssa> lmk if that fixes it
<alyssa> ok yes that should fix it
<alyssa> change PIPE_MAX_SAMPLERS to nr_samplers in line 1351
<alyssa> bind_sampler_states doesn't zero [nr_samplers, PIPE_MAX_SAMPLERS) so can get non-null garbage samplers leading to a use-after-free (segfault derefencing sampler) or a buffer overflow (writing samplers[] out of bounds, we only allocated room for nr_samplers of them) or both
<alyssa> thank you for the helpful bug report :-)
* alyssa really appreciates bug reports that she can actually do something about
<alyssa> "Reproducible segfault in this workload, here is the backtrace, here is the line number that's failing, I am using the stable tree with no patches" -- good
<alyssa> "This workload used to work on the stable tree, but it no longer works on the stable tree. I've bisected the failure down to commit sha 123456. If I revert that commit on top of main, it works again." -- good
<alyssa> "I just tried this random DirectX game and it doesn't work very good, here is an apitrace, please fix, cc @someone_banned_for_code_of_conduct_violations" -- not good
SSJ_GZ has quit [Ping timeout: 480 seconds]
<Dcow> yeah, that's do it. Thank you!
mps has left #asahi-gpu [#asahi-gpu]
<Dcow> sorry for delay, wanted to _test_ the whole match >_>
<Dcow> lol xonotic have a good perf even under gdb!
<Dcow> I should go sleep or I miss start of l-i-n-a's stream again
Cyrinux has joined #asahi-gpu
Dcow has quit [Quit: dcow]
<alyssa> +1, thanks!
kenzie7 has quit []
kenzie7 has joined #asahi-gpu